On Sun, Aug 16, 2020 at 08:47 PM, Roger Need wrote:
Glen ,
After reading through all the posts I see you did not get an answer on where
you could BUY reasonably priced cal loads for your NanoVNA. Here is a link to
SDR-Kits which has cal loads for SMA, BNC and N connectors in different
quality and price ranges.
SMA -
BNC -
N -
Roger
They may be reasonably priced, but don't count on them being any better. I have the female Rosenberger SMA kit and the load has a DC resistance of 48.82R. When you look at the specs, this should not be surprising, Rosenberger guarantees 32.2dB DC-2GHz and 26.4dB 2-8GHz. 48.82R is >37dB, so there is a good chance it is still in spec (and probably good enough for most purposes). Note that this is typical for a general purpose termination. Don't expect much tighter specifications on anything that is not sold as a calibration standard (and priced accordingly). Typical performance will be better, but unless you have something better to compare it to (not just at DC), you won't really know for sure.
A side note: if I correct for the DC resistance, the Rosenberger load is within <-45dB (@3GHz) of the other female SMA load I have (Huber&Suhner 18GHz). This might just be luck and I don't know how good the other load really is (at DC it happens to be almost exactly 50.0R). Still, at lower frequencies, just correcting for DC resistance during calibration (or selecting the load for DC resistance) might actually work quite well. If the reactive component were already dominant at a couple of 100 MHz, there is probably not much chance of meeting the spec at 18GHz... For some random home-made termination (or ebay special) things will be different. Of course, at higher frequencies this will eventually fall apart as well. Unless you know the offset delay as well and have your reference plane set properly, you might actually make things worse.
As others have mentioned, measuring the DC resistance acurately is not totally straightforward either. A normal 3.5 digit multimeter may easily be 0.5%+several digits on the resistance range. Something like the 34401A I used should be within +/- 0.02R. A high-quality banana->BNC and a BNC->SMA adapter ended up working fine, but the cheap ones I tried initially didn't give repeatable results (the contact resistance varied wildly). 4-wire is not necessary, just zeroing with a short is fine (and with the good adapters proved not necessary at all). If I had to do this on a regularly, I would probably make a special fixture. But this was just a one-time experiment out of curiosity.
